JOURNAL OF NATURAL RESOURCES ›› 2016, Vol. 31 ›› Issue (11): 1906-1917.doi: 10.11849/zrzyxb.20151336

Previous Articles     Next Articles

Modeling Extreme Precipitation in the Poyang Lake Basin Based on Stationary and Non-stationary GEV Models

YIN Yi-xing1a,1b, CHEN Hai-shan1a, XU Chong-yu2, CHEN Ying3, ZHAO Jun1b, SUN Shan-lei1a   

  1. 1. a. China Key Laboratory of Meteorological Disaster of Ministry of Education, b. College of Hydrometeorology, Nanjing University of Information Science and Technology, Nanjing 210044, China;
    2. Department of Geosciences, University of Oslo, Oslo, Norway;
    3. College of Geographic Sciences, Fujian Normal University, Fuzhou 350007, China
  • Received:2015-12-04 Revised:2016-04-18 Online:2016-11-20 Published:2016-11-20
  • Supported by:
    National Natural Science Foundation of China, No.41671022; Natural Science Foundation of the Jiangsu Higher Education Institutions,No.15KJB170014; China Postdoctoral Science Foundation Funded Project, No.2013M531384; Jiangsu Planned Projects for Postdoctoral Research Funds, No.1301136C

Abstract: This paper chooses precipitation data of 14 meteorological stations in the Poyang Lake Basin to model the changes of extreme precipitation from 1951 to 2010 based on both stationary and non-stationary Generalized Extreme Value (GEV) models. The non-stationary characteristics of the annual maximum 1-day series (AMS1) were detected, and the non-stationary GEV model which used the time as a covariate to the location parameter is utilized for non-stationary AMS1. The results showed that: 1) The shape parameters of the AMS1 are all bigger than 0, and follow the Fréchet distribution; the spatial distribution of the GEV location and scale parameters are quite consistent with each other, but the spatial distribution of shape parameters is not consistent with them. 2) The confidence interval given by Profile methods are more accurate for longer return periods in comparison to Delta method, and evident asymmetry appears in the return level’s Profile log-likelihood curve for longer rerun periods. 3) The spatial pattern of the extreme precipitation for different return periods are obtained, and the patterns are in line with the patterns of the location and scale parameters, but are different from the pattern of shape parameters. 4) The time-varying return levels (effective return level) for different return periods are obtained from the non-stationary GEV model of Ganxian Station. The return level of 100 years in 1951 will decrease to 50 years in 2010 for AMS1, which indicates greater risk for extreme precipitation and flood disasters in the future.

Key words: extreme precipitation, GEV model, non-stationary, Poyang Lake Basin

CLC Number: 

  • TV125